On 6/15/21 1:52 PM, Tom Lane wrote:
> Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
>> On 2021-Jun-15, Tom Lane wrote:
>>> I think this ought to be reverted and reviewed more carefully.
>
>> It seems to me that removing the cast-to-range[] is a sufficient fix,
>> and that we can do with only the unnest part for pg14; the casts can be
>> added in 15 (if at all). That would mean reverting only half the patch.
>
> Might be a reasonable solution. But right now I'm annoyed that the
> buildfarm is broken, and I'm also convinced that this didn't get
> adequate testing.
I had focused testing primarily on the "unnest" cases that I had
described in my original note. I did a couple of casts and had no issue;
I did not test with pg_dump / pg_upgrade, but noting to do so in the
future in cases like this.
> I think "revert and reconsider" is the way
> forward for today.
I don't want the buildfarm broken so I'm fine if this is the best way
forward. If we can keep the "unnest" functionality I would strongly
suggest it as that was the premise of the original note to complete the
utility of multiranges for v14. The casting, while convenient, is not
needed.
Jonathan